WO2018076765A1

WO2018076765A1 - Content distribution method and device for cloud computing system, computing node and system

Info

Publication number: WO2018076765A1
Application number: PCT/CN2017/091276
Authority: WO
Inventors: 刘成华; 肖腾飞
Original assignee: 华为技术有限公司
Priority date: 2016-10-31
Filing date: 2017-06-30
Publication date: 2018-05-03
Also published as: CN108023812A; CN108023812B

Abstract

Provided are a content distribution method and device for a cloud computing system, a computing node and a system. The method comprises: receiving a content distribution instruction, the content distribution instruction comprising an identifier of content to be distributed and a distribution target node list, and the distribution target node list comprising an identifier of at least one distribution target node; according to physical networking topology, the state traffic of each node in the system and the distribution target node list, generating a content distribution path, the content distribution path being a one-way chain distribution path composed of a head node and at least one distribution target node; and sending a content distribution message to the at least one distribution target node. This method can ensure that the number of nodes connected to each node in the content distribution path is controlled by a distribution controller; therefore, the balance of the access bandwidth of each distribution target node can be realized. Moreover, it can also be ensured that the content distribution path generated each time content is distributed is the optimal distribution path.

Description

Content distribution method and device, computing node and system of cloud computing system

The present application claims priority to Chinese Patent Application No. 201610931630.X filed on October 31, 2016, entitled "Content Distribution Method and Apparatus, Computing Node and System of Cloud Computing System", The entire contents are incorporated herein by reference.

Technical field

The present invention relates to communications technologies, and in particular, to a content distribution method and apparatus, a computing node, and a system of a cloud computing system.

Background technique

Cloud computing originated from large Internet companies, which refers to the delivery and use of IT infrastructure. Through the network, the required resources such as hardware, platform, software, etc. are obtained in an on-demand and scalable manner. Virtualization technology is the main supporting technology in the field of cloud computing. A cloud computing center includes multiple physical servers. Each physical server can be divided into multiple virtual machines, each with independent computing and processing capabilities. In a cloud computing system, there may be scenarios where the same content is distributed to a large number of virtual machines. For example, distribute the same operating system patch for many virtual machines. In general, the number of physical servers in a cloud computing system can reach 100,000 units. Correspondingly, the number of virtual machines in a cloud computing system is even larger, generally reaching hundreds of thousands or millions. In this case, How to make a large number of virtual machines download to the distributed content quickly is an important issue that the cloud computing system needs to solve when distributing content.

In the prior art, peer-to-peer (P2P) technology is used to improve download speed. P2P is a distributed system without centralized control. Each node in the system is both a client and a server. The BitTorrent (BT) distribution technology is a typical P2P distribution technology. Specifically, the BT system includes a central index server and a plurality of nodes. The central index server receives and counts information such as the IP address and port number of all downloaders. And sending a list of nodes to each download terminal, the download terminal can establish a connection with the node in the node list, and download data from the node that establishes the connection. The download terminal can periodically connect to the central index server to inform the central index server of the content that it has downloaded. That is, in the BT system, each node can establish a connection with a plurality of other nodes, and download data from other connected nodes, and requires the node to occupy as much bandwidth as possible when downloading data, so as to improve downloading. speed.

However, the prior art has an unpredictable problem of system traffic, and therefore cannot meet the requirements of the cloud computing system for system flow control.

Summary of the invention

The embodiments of the present invention provide a content distribution method and device, a computing node, and a system for a cloud computing system, which are used to solve the problem that the system traffic in the prior art in the content distribution of the cloud computing system is unpredictable.

A first aspect of the embodiments of the present invention provides a content distribution method of a cloud computing system, which is described in the perspective of a distribution controller, and the method includes:

First, a content distribution instruction is included, where the content distribution indication includes an identifier of the content to be distributed and a distribution target node list, and the distribution target node list includes an identifier of the at least one distribution target node.

Secondly, a content distribution path is generated according to a state traffic of each node in the physical networking topology cloud computing system of the cloud computing system and a distribution target node list, wherein the content distribution path is performed by the first node and at least one distribution target node A unidirectional chain distribution path is formed, and the head node is a node in the cloud computing system having a complete content to be distributed.

Further, the content distribution message is sent to the at least one distribution target node, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, so that the at least one distribution target node acquires and distributes according to the content distribution path and the identifier of the content to be distributed. Content to be distributed.

When the content distribution is performed, the content distribution path is generated according to the physical networking topology of the cloud computing system, the state traffic of each node in the cloud computing system, and the distribution target node list, and the content distribution is performed according to the content distribution path. First, the content distribution path is generated according to the physical networking topology information of the cloud computing system and the state traffic of each node in the cloud computing system, so that it is ensured that there is no disordered cross-connection between the distribution target nodes in the content distribution path. And ensuring that the actual traffic of each distribution target node in the content distribution path is within a controllable range. Secondly, the content distribution path is a unidirectional chain path, that is, the number of nodes connected to each node in the content distribution path is controlled by the distribution controller, so that the balance of the inbound and outbound bandwidth of each distribution target node can be achieved. Therefore, the method can ensure that the selected content distribution path is an optimal content distribution path. In addition, the method can generate a content distribution path according to the physical networking topology information of the cloud computing system and the state traffic of each node in the cloud computing system each time the content distribution instruction is received, thereby ensuring that the content is performed each time. The content distribution paths generated at the time of distribution are the optimal distribution paths. Further, based on the above-described optimal content distribution path, the flow control in the content distribution process can be fully realized.

In a possible design, the content distribution message further includes a specified distribution rate of the at least one distribution target node, which is determined according to the state traffic of the at least one distribution target node when the content distribution path is generated.

In one possible design, the method further includes:

Receiving a ready state reported by the at least one distribution target node, and transmitting a content receiving and forwarding indication to the at least one distribution target node according to the ready state.

In one possible design, when generating a content distribution path, it can be generated in the following ways:

According to the physical networking topology and the distribution target node list, two distribution target nodes that are located on the same physical host or the same chassis in the physical networking are used as adjacent nodes in the content distribution path;

Determining a node in the cloud computing system that has a complete content to be distributed;

According to the state traffic of each node in the cloud computing system, the node with the lowest current traffic consumption among the nodes having the complete content to be distributed is used as the first node in the content distribution path.

In a possible design, when there are multiple nodes in the cloud computing system with complete content to be distributed, the content distribution path can be generated by the following methods:

Determining a plurality of nodes in the cloud computing system that have complete content to be distributed;

Generating a plurality of content distribution paths, wherein the number of generated content distribution paths is less than or equal to the number of nodes having the complete content to be distributed, and the first node in each of the content distribution paths is One of the nodes having the complete content to be distributed.

In one possible design, it also includes:

Setting a preset number of nodes in the content distribution path;

If the number of nodes in the content distribution path is greater than the preset number of nodes, selecting a fork node from the content distribution path, and splitting the content distribution path into the first content according to the fork node a distribution path and a second content distribution path;

The first content distribution path and the first node of the second content distribution path are the same, and the first content distribution path and the second content distribution path are demarcated by the bifurcation node, and the bifurcation The node is the node with the lowest traffic consumption in the content distribution path.

In one possible design, it also includes:

Determining a faulty node in the content distribution path;

And deleting the faulty node from the content distribution path, and using a downstream node of the faulty node as a downstream node of the upstream node of the faulty node.

In one possible design, the first way to determine the faulty node in the content distribution path is:

If the response message to the content distribution message sent by the first distribution target node is not received, determining that the first distribution target node is a faulty node in the content distribution path, where the first distribution target The node distributes the target node for any of the content distribution paths.

In a possible design, the second way of determining the faulty node in the content distribution path is as follows:

Determining a faulty node in the content distribution path according to the status traffic reported by the at least one distribution target node and the distribution log.

In one possible design, the third way to determine the faulty node in the content distribution path is:

If the distribution failure message sent by the second distribution target node is received, determining that the downstream node of the second distribution target node is one of the content distribution paths, wherein the second distribution target node is the Any one of the distribution target nodes except the last distribution destination node in the content distribution path.

In a possible design, the above method further includes:

Acquiring a recovery node in the content distribution path, the recovery node being a node deleted from the content distribution path due to a failure;

Adding the recovery node to the content distribution path, so that the recovery node acquires distribution content from an upstream node of the recovery node, wherein the recovery node is from an upstream node of the recovery node by an active request mode Get the distribution.

In a possible design, the above method further includes:

The method also includes:

Receiving a distribution progress, a distribution log, a distribution result, and a status flow periodically reported by the at least one distribution target node;

The distribution progress, the distribution log, the distribution result, and the status traffic are transmitted to the terminal.

In a possible design, when receiving the distribution progress, the distribution log, the distribution result, and the status traffic periodically reported by the distribution target node, the HTTPS port is used to receive the distribution progress and the distribution log periodically reported by the distribution target node. , distribution results, and status traffic.

In one possible design, a content distribution message is sent to the distribution target node using an HTTPS port when the content distribution message is sent to the distribution target node.

A second aspect of the embodiments of the present invention provides a content distribution method of a cloud computing system, where the method is described in the perspective of a computing node, and the method includes:

First, a content distribution message is received, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, wherein the content distribution path is determined by the distribution controller according to a physical networking topology of the cloud computing system, and a status of each node in the cloud computing system. The traffic is generated by the distribution target node list, and the content distribution path is a unidirectional chain distribution path formed by the first node and the at least one distribution target node, and the first node is a node in the cloud computing system that has the complete content to be distributed.

Secondly, according to the identifier of the content to be distributed, the metadata of the content to be distributed is obtained from the head node, and the metadata of the content to be distributed includes a digital signature, a block index, and a block hash value of the content to be distributed.

Further, the content distribution processing is performed based on the content distribution path and the metadata of the content to be distributed.

In a possible design, the content distribution message further includes a specified distribution rate of the distribution target node; correspondingly, when the content distribution processing is performed according to the content distribution path and the metadata of the content to be distributed, the content distribution path may be And the metadata of the content to be distributed, and the content distribution processing is performed at a specified distribution rate.

In one possible design, the method further includes:

Reporting the ready status to the distribution controller;

Receiving a content receiving and forwarding indication sent by the distribution controller.

In a possible design, when the content distribution processing is performed according to the content distribution path and the metadata of the content to be distributed, the following may be specifically performed by:

Receiving a block sent by an upstream node;

And performing a hash check on the block according to the block index and the block hash value in the metadata of the content to be distributed, and acquiring a verification result;

If the verification result is that the verification is successful, the block is saved, and the block is sent to a downstream node in the content distribution path according to the specified distribution rate.

In a possible design, when the block is sent to the downstream node in the content distribution path according to the specified distribution rate, the specific one may be:

If the transmission fails, resend it;

If the number of retransmission failures reaches a preset number of times, a distribution failure message is sent to the distribution controller to cause the distribution controller to re-establish the content distribution path according to the state of the downstream node.

In one possible design, it also includes:

If all the blocks corresponding to the content to be distributed are received, the digital signature is checked according to the digital signature, and the result of the digital signature check is reported to the distribution controller.

In one possible design, it also includes:

Obtain distribution progress, distribution logs, distribution results, and status traffic;

The distribution progress, distribution log, distribution result, and status traffic are reported to the distribution controller.

In a possible design, when the distribution progress, the distribution log, the distribution result, and the status traffic are reported to the distribution controller, the distribution progress, the distribution log, and the distribution log may be reported to the distribution controller by using an HTTPS protocol. Distribute results and status traffic.

A third aspect of the embodiments of the present invention provides a content distribution method, which is described in the home of a content repository, and the method includes:

Receiving a creation metadata indication, where the creation metadata indication includes an identifier of the content to be distributed;

Metadata of the content to be distributed is generated according to the identifier of the content to be distributed.

In a possible design, when the metadata of the content to be distributed is generated according to the identifier of the content to be distributed, the method may be specifically performed by the following method:

Determining whether the content to be distributed exists according to the content identifier to be distributed, and if yes,

Performing partial division and block division on the to-be-distributed content, acquiring part number, partial size, block size, and block index; and,

Digitally signing the content to be distributed to obtain a digital signature; and,

Hashing each block of the partition to obtain a block hash value of each block;

The partial number, the size of each part, the block size, the block index, the digital signature, and the block hash value of each block are written into the metadata of the content to be distributed.

A fourth aspect of the present invention provides a content distribution apparatus, including:

a receiving module, configured to receive a content distribution indication, where the content distribution indication includes an identifier of the content to be distributed and a distribution target node list, where the distribution target node list includes an identifier of the at least one distribution target node;

a processing module, configured to generate a content distribution path according to a physical networking topology of the cloud computing system, a state traffic of each node in the cloud computing system, and the distribution target node list, where the content distribution path is a first node and a a unidirectional chain distribution path formed by at least one distribution target node, where the first node is a node having a complete content to be distributed in the cloud computing system;

a sending module, configured to send a content distribution message to the at least one distribution target node, where the content distribution message includes the content distribution path and an identifier of the content to be distributed, so that the at least one distribution target node is configured according to the The content distribution path and the identifier of the content to be distributed acquire and distribute the content to be distributed.

In a possible design, the content distribution message further includes a specified distribution rate of the at least one distribution target node, where the specified distribution rate is based on the at least one distribution target node when the content distribution path is generated The status of the traffic is determined.

In a possible design, the receiving module is further configured to receive a ready state reported by the at least one distribution target node;

The sending module is further configured to send a content receiving and forwarding indication to the at least one distribution target node according to the ready state.

In one possible design, the processing module is further configured to:

According to the physical networking topology and the distribution target node list, two distribution target nodes located on the same physical host or the same chassis in the physical networking are used as adjacent nodes in the content distribution path;

Determining a plurality of nodes in the system having the complete content to be distributed; and,

In one possible design, the processing module is further configured to:

Setting a preset number of nodes in the content distribution path; and,

In one possible design, the processing module is further configured to:

Determining a faulty node in the content distribution path; and,

In a possible design, the processing module further includes:

a fourth determining unit, configured to determine, when the response message for the content distribution message sent by the first distribution target node is not received, the first distribution target node is a faulty node in the content distribution path, where The first distribution target node is a distribution target node of any one of the content distribution paths.

In a possible design, the processing module further includes:

And a fifth determining unit, configured to determine a faulty node in the content distribution path according to the state traffic reported by the at least one distribution target node and the distribution log.

In a possible design, the processing module further includes:

a sixth determining unit, configured to: when receiving the distribution failure message sent by the second distribution target node, determine that the downstream node of the second distribution target node is a fault node in the content distribution path, where the The second distribution target node is any one of the distribution target nodes except the last distribution destination node in the content distribution path.

In one possible design, the processing module is further configured to:

In a possible design, the receiving module is further configured to receive a distribution progress, a distribution log, a distribution result, and a status flow periodically reported by the at least one distribution target node;

The sending module is further configured to send the distribution progress, the distribution log, the distribution result, and the status traffic to the terminal.

A fifth aspect of the embodiments of the present invention provides a content distribution apparatus, including:

a receiving module, configured to receive a content distribution message, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, wherein the content distribution path is configured by the distribution controller according to a physical networking topology of the cloud computing system, and the cloud computing a state traffic of each node in the system and a distribution target node list, where the content distribution path is a unidirectional chain distribution path formed by the first node and the at least one distribution target node, where the first node is in the system a complete node of the content to be distributed;

a processing module, configured to acquire, according to the identifier of the content to be distributed, metadata of the content to be distributed from a first node, where the metadata of the content to be distributed includes a digital signature, a block index, and a content of the content to be distributed Block hash value;

The processing module is further configured to perform content distribution processing according to the content distribution path and the metadata of the content to be distributed.

In a possible design, the content distribution message further includes a specified distribution rate of the distribution target node;

The processing module is specifically configured to perform content distribution processing according to the specified distribution rate according to the content distribution path and the metadata of the content to be distributed.

In one possible design, it also includes:

a sending module, configured to report a ready state to the distribution controller;

The receiving module is further configured to receive a content receiving and forwarding indication sent by the distribution controller.

In one possible design, the processing module includes:

a receiving unit, configured to receive a block sent by an upstream node;

a checking unit, configured to perform hash check on the block according to the block index and the block hash value in the metadata of the content to be distributed, and obtain a verification result;

And a sending unit, configured to save the block when the verification result is a verification success, and send the block to a downstream node in the content distribution path according to the specified distribution rate.

In a possible design, the sending unit is specifically configured to:

If the transmission fails, resend; and,

In a possible design, the processing module is further configured to: when receiving all the blocks corresponding to the content to be distributed, perform digital signature check on the to-be-distributed content according to the digital signature, and digitally sign the The result of the check is reported to the distribution controller.

In a possible design, the processing module is further configured to obtain a distribution progress, a distribution log, a distribution result, and a status flow;

The sending module is further configured to report the distribution progress, the distribution log, the distribution result, and the status traffic to the distribution controller.

A sixth aspect of the embodiments of the present invention provides a content distribution apparatus, including:

a receiving module, configured to receive a creation metadata indication, where the creation metadata indication includes an identifier of the content to be distributed;

And a processing module, configured to generate metadata of the content to be distributed according to the identifier of the content to be distributed.

In one possible design, the processing module includes:

a dividing unit, configured to perform partial division and block division on the to-be-distributed content after determining that the content to be distributed exists according to the content identifier to be distributed, and obtain a partial quantity, a partial size, a block size, and a block index.

The processing unit is configured to perform digital signature processing on the content to be distributed, and obtain a digital signature.

A check unit is configured to perform hash check on each of the divided blocks to obtain a block hash value of each block.

A write unit for writing a partial number, a partial size, a block size, a block index, a digital signature, and a block hash value of each block into the metadata of the content to be distributed.

A seventh aspect of the embodiments of the present invention provides a distribution controller, including:

Memory and processor.

The memory is used to store program instructions, and the processor is used to call program instructions in the memory to perform the following methods:

Receiving a content distribution indication, where the content distribution indication includes an identifier of the content to be distributed and a distribution target node list, where the distribution target node list includes an identifier of the at least one distribution target node;

Generating a content distribution path according to a physical networking topology of the cloud computing system, a state traffic of each node in the cloud computing system, and the distribution target node list, wherein the content distribution path is a first node and the at least one distribution target a unidirectional chain distribution path formed by the node, where the head node is a node having a complete content to be distributed in the cloud computing system;

And transmitting, to the at least one distribution target node, a content distribution message, where the content distribution message includes the content distribution path and an identifier of the content to be distributed, so that the at least one distribution target node according to the content distribution path and The identifier of the content to be distributed acquires and distributes the content to be distributed.

An eighth aspect of the embodiments of the present invention provides a computing node, including:

Memory and processor.

Receiving a content distribution message, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, wherein the content distribution path is configured by the distribution controller according to a physical networking topology of the cloud computing system, and each node in the cloud computing system The status flow is generated by the distribution target node list, and the content distribution path is a unidirectional chain distribution path formed by the first node and the at least one distribution target node, where the first node is a complete system in the cloud computing system. a node that describes the content to be distributed;

Obtaining, according to the identifier of the content to be distributed, the metadata of the content to be distributed from the first node, where the metadata of the content to be distributed includes a digital signature, a block index, and a block hash value of the content to be distributed;

Content distribution processing is performed based on the content distribution path and metadata of the content to be distributed.

A ninth aspect of the embodiments of the present invention provides a content repository, including:

Memory and processor.

Receiving a creation metadata indication, wherein the creation metadata indication includes an identifier of the content to be distributed;

Generating metadata of the content to be distributed according to the identifier of the content to be distributed.

A tenth aspect of the present invention provides a content distribution system including the foregoing distribution controller, a computing node, and a content repository.

In the background art, for a content distribution, there is no determined distribution path, but the distribution target node dynamically establishes a connection with other nodes in the process of downloading. In order to exchange content more effectively, one distribution target node may establish a large number of The connection may cause a sudden burst of traffic, so the background technology cannot achieve effective flow control. Secondly, when downloading a file block, it is not downloaded or uploaded in order. Before downloading, first write the file with the same file capacity on the hard disk, and then randomly write the data by filling method. Each download needs to perform check calculation for all blocks. The read and write operations are performed simultaneously during the download and upload process, and the hard disk resources are occupied. Therefore, the efficiency of content distribution cannot be guaranteed. In addition, the topology construction in the background art requires the support of the index server, and the order in which the nodes join the system has an extremely important influence on the formation of the topology. When there is a node to join, the download and upload can be started, and if the node is downloaded, the exit may occur. The number of "nodes" is zero, which causes some nodes to slow down and even download all the file blocks. Therefore, the background technology cannot guarantee the integrity of the distributed content.

The solution provided by the embodiment of the present invention can implement flow control in the content distribution process compared to the prior art.

DRAWINGS

In order to more clearly illustrate the present invention or the technical solutions in the prior art, a brief description of the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are Some of the embodiments of the invention may be obtained by those of ordinary skill in the art in view of the drawings without departing from the scope of the invention.

1 is a system architecture diagram of a content distribution method according to an embodiment of the present invention;

2 is an interaction flowchart of Embodiment 1 of a content distribution method according to an embodiment of the present invention;

3 is an example of a central physical server of a cloud computing system;

4 is an example of internal connection of a central physical server of a cloud computing system;

Figure 5 is a schematic diagram of a single chain path;

6 is a schematic diagram of parallel distribution of multiple content distribution paths;

7 is a schematic diagram of a content distribution path bifurcation;

8 is a schematic diagram of deleting a faulty node from a content distribution path;

9 is a schematic diagram of rejoining a recovery node to a content distribution path;

Figure 10 is a schematic diagram of content distribution to multiple subnets;

11 is a schematic flowchart of generating metadata according to Embodiment 2 of the present invention;

12 is a schematic diagram of hierarchical division of content to be distributed;

Figure 13 is a diagram showing an example of metadata of content to be distributed;

FIG. 14 is a schematic diagram of a content receiving and forwarding process according to Embodiment 3 of a content distribution method according to an embodiment of the present disclosure;

FIG. 15 is a block diagram of a first embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 16 is a block diagram of a second embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 17 is a block diagram of a third embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 18 is a block diagram of a fourth embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 19 is a block diagram showing a structure of a fifth embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 20 is a block diagram of a first embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 21 is a block diagram of a second embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

Figure 22 is a block diagram showing the structure of a third embodiment of a content distribution apparatus according to an embodiment of the present invention;

FIG. 23 is a block diagram of a first embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 24 is a block diagram of a second embodiment of a content distribution apparatus according to an embodiment of the present disclosure;

FIG. 25 is a physical block diagram of Embodiment 1 of a distribution controller according to an embodiment of the present disclosure;

FIG. 26 is a physical block diagram of Embodiment 1 of a computing node according to an embodiment of the present disclosure;

FIG. 27 is a physical block diagram of Embodiment 1 of a content repository according to an embodiment of the present invention.

detailed description

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. Some embodiments, rather than all of the embodiments, are invented. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

Before describing the embodiments of the present invention in detail, first, several concepts involved in the embodiments of the present invention are explained:

1. System, node, and physical networking topology

The system involved in the embodiments of the present invention is a system consisting of a group of mutually independent computing nodes interconnected by a high-speed network, and the same content information needs to be shared or distributed between the computing nodes. The deployment form of the system may be distributed deployment and centralized deployment.

For example, the system can be a cloud computing system. Each virtual machine in the cloud computing system can be regarded as a computing node in the system, and the same content information needs to be shared or distributed between the virtual machines.

It should be noted that the system according to the embodiment of the present invention is not limited to a specific system, and any system in which the internal computing node needs to share or distribute the same content information may use the embodiment provided by the embodiment of the present invention. method.

For convenience of description, the “computation node” is collectively referred to as “node” in the embodiment of the present invention.

For each system, it corresponds to a physical network. The physical network topology of the system refers to the location and connection relationship between physical devices in the physical network corresponding to the system.

2. Adjacent nodes, upstream nodes, downstream nodes

In the embodiment of the present invention, the adjacent nodes refer to two nodes that are adjacent in the content distribution path, and the node 1, the node 2, and the node 3 in FIG. 1 are taken as an example, and the distribution path 1 in FIG. Node 1, Node 2, and Node 3, wherein Node 2 receives the distribution content from Node 1, that is, Node 2 and Node 1 are adjacent in the distribution path 1, and therefore, Node 1 and Node 2 are adjacent nodes. At the same time, node 3 and node 2 are also adjacent nodes.

For a particular node 2, node 1 is its upstream node and node 3 is its downstream node, ie, for a particular node 2, it receives the distribution content from the upstream node and sends the distribution to the unloading node.

3, traffic and bandwidth

Traffic refers to the amount of distributed data that the node uplinks (transmits) or downlinks (receives), that is, the amount of data sent or received by the node over a period of time. A node with a low current traffic rate refers to a node that transmits and receives less data during the current statistical time period.

Bandwidth refers to the transmission speed allowed by the maximum data of the node within a specified time. The flow control is achieved by limiting the bandwidth occupied by the node distribution process, the number of uplink and downlink connections of the node, and performing differentiated forwarding and packet loss control. of.

FIG. 1 is a system architecture diagram of a content distribution method according to an embodiment of the present invention. As shown in FIG. 1, the system includes a distribution controller, a content warehouse, and a plurality of nodes.

The distribution controller creates a content distribution path according to the information of the status, the traffic, the physical location, and the like of the plurality of nodes, and controls the plurality of nodes. In the embodiment of the present invention, the distribution controller and the content distribution device may be the same. The content repository is used to store the content to be distributed, and the content to be distributed is sent to the head node in the distribution path. Each content distribution path is a unidirectional chain distribution path, and the first node in each distribution path is a node including the complete distribution content, and the first node starts to distribute the content to the nodes in the distribution path in order. In a embodiment of the present invention, when a user needs to distribute the content to a distribution target node in the cloud computing system, the distribution controller receives a content distribution indication sent by the terminal that needs to distribute the content, where the content distribution indication includes the content to be distributed. And identifying a distribution target node list, wherein the distribution target node list includes an identifier of at least one distribution target node; at this time, the distribution controller determines which nodes the content needs to be distributed to. The distribution controller generates a content distribution path according to the physical networking topology of the cloud computing system, the state traffic of each node, and the distribution target node list; and then sends a content distribution message including the content distribution path and the identifier of the content to be distributed to each distribution target. node. Each node receives the content distribution path, establishes a content link channel with the upstream and downstream nodes according to the content distribution path, and starts content distribution. A specific content distribution process of an embodiment of the present invention is shown in FIG. 2.

FIG. 2 is an interaction flowchart of a content distribution method according to Embodiment 1 of the present invention. The interaction process may utilize the system architecture of FIG. 1 and the network element therein to implement content distribution of the cloud computing system, as shown in FIG. 2 . As shown, the method includes:

S201. The user issues content to the content warehouse through the terminal.

Specifically, when there is new content that can be distributed, for example, after a new operating system patch is newly generated, the user can publish the new distributable content to the content repository through the terminal. The terminal can directly publish the content to the content repository by using an Application Programming Interface (API) provided by the content repository. When the content is published, the content is saved in the content repository.

The content repository stores the distributed content, and the content repository can generate a content identifier for the content, the content identifier uniquely identifies the content, and returns the content identifier to the terminal. Subsequent when the terminal or distribution controller needs to obtain content from the content repository, it needs to distinguish by the content identifier.

S202. The user sends a content distribution indication to the distribution controller through the terminal.

The content distribution indication includes an identifier of the content to be distributed and a distribution target node list, and the distribution target node list includes an identifier of the at least one distribution target node.

Specifically, when the user wishes to distribute a certain content that has been posted to the content repository to a specific target node, the content distribution indication is sent to the distribution controller through the terminal, and the content distribution indication indicates the identification and distribution of the content to be distributed. List of target nodes.

The identifier of the content to be distributed is used to indicate which content in the content warehouse is specifically distributed by the distribution controller, and the identifier of the content to be distributed is the identifier allocated by the content warehouse in the step S201.

In addition, the distribution target node list includes an identifier of at least one distribution target node, and the distribution target node refers to that the user wants to distribute the content to the destination node, and the nodes that need to be distributed to different contents may not be the same, and therefore, the user passes Each time the terminal sends a content distribution indication to the distribution controller, it needs to specify the target node to which the content is specifically sent. The number of the distribution target nodes may be one or multiple, and the present invention does not limit the number thereof. The identity of each distribution target node can be represented by the name, number, or IP address of the node, and the like.

It should be noted that the above steps S201 and S202 may be discontinuous in time, and the user may first publish the content to the content repository, and then select an appropriate time to send the content distribution indication to the distribution controller.

Optionally, before sending the content distribution indication to the distribution controller, the terminal may first send an inquiry request to the content repository according to the identifier of the content to be distributed, and the content warehouse may determine whether the content corresponding to the identifier exists after receiving the query request, if If yes, the terminal returns a query success message to the terminal, and the terminal may continue to send the content distribution indication to the distribution controller. If not, the terminal may return a failure message to the terminal, and the terminal may only prompt the user for the content corresponding to the identifier, and prompt the user to re-confirm the content. Identify, without continuing to send a content distribution indication to the distribution controller.

S203. The distribution controller generates a content distribution path according to the content distribution instruction.

After receiving the content distribution indication sent by the terminal, the distribution controller generates a content distribution path according to the physical networking topology, the status traffic of each node in the system, and the distribution target node list carried in the content distribution indication. The physical networking topology indicates the current networking connection relationship between system nodes. The status traffic of each node in the system indicates the status traffic status of each node. For example, the node status traffic can be the specific current traffic load situation, or the node. Whether the current traffic exceeds the threshold. Each time the distribution controller receives the content distribution instruction, it combines these two actual situations to select a content distribution path that can save the distribution time and save system traffic, thereby ensuring efficient content distribution and realizing flow control.

The content distribution path is a unidirectional chain distribution path formed by the first node and the at least one distribution target node, and the first node in the content distribution path is the first node, and the node with complete content to be distributed in the system can be the first node.

Specifically, the content distribution path generated by the distribution controller includes a first node and at least one distribution target node, and the connection relationship between the nodes is unidirectional chain, for example, for a node in the content distribution path. A, it can only receive data from one Node B, and can only send the received data to one node C, and can no longer send data to Node B, nor can it receive data from Node C.

In addition, in order to improve the efficiency of content distribution, content distribution can be used in the distribution process, that is, a complete distribution of content to be distributed is divided into multiple blocks, multiple pieces are distributed one by one, and finally combined into one complete content. During the content distribution process, some nodes in the system may have all the blocks corresponding to the content to be distributed, that is, the complete content to be distributed, and these nodes can be the first nodes in the content distribution path. Some nodes may have a part of the blocks, and these nodes cannot be the first node in the content distribution path.

The distribution controller generates a content distribution path based on the physical networking topology and the state traffic of each node in the system. A chained path, that is, the number of nodes to which each node in the content distribution path is connected can be controlled by a distribution controller. Specifically, it may be in the form of a single chain path, a plurality of concurrent chain paths, etc., and different types of chain paths may be used in different scenarios, which will be described in detail later in the embodiments of the present invention.

S204. The distribution controller sends a creation metadata indication to the content warehouse. The creation metadata indication includes an identifier of the content to be distributed.

Specifically, after the distribution controller generates the content distribution path, the distribution controller may send a creation metadata indication to the content repository, the creation metadata indicating metadata for instructing the content repository to establish the content to be distributed, the content to be distributed The metadata is the metadata formed after the content to be distributed and the security check is performed. In the content forwarding process, it is necessary to perform security according to the metadata and will be described in detail later.

Since the first node has complete content to be distributed, that is, the first node has used metadata in the content forwarding process, metadata of the content to be distributed is also required in the first node.

S205. The content repository generates metadata of the content to be distributed according to the creation metadata indication sent by the distribution controller.

Specifically, after receiving the content distribution instruction, the content repository determines the to-be-distributed content saved in the content repository according to the identifier of the content to be distributed carried therein, and generates metadata of the content to be distributed. Before generating the metadata of the content to be distributed, the content repository first determines whether the metadata of the content to be distributed has been generated, and if it has been generated, the generation operation is no longer performed, but the existing metadata can be directly used.

The metadata of the content to be distributed includes a digital signature of the content to be distributed, a block index, a block hash value, and the like, and a method of specifically generating metadata of the content to be distributed will be explained in detail in the introduction of FIG. 11 below.

S206. The distribution controller sends a content distribution message to all distribution target nodes in the content distribution path.

2, only the distribution controller sends a content distribution message to the two distribution target nodes, node 1 and node 2, but in the actual distribution process, each content distribution path may include multiple distribution target nodes, and distribute The controller can send a content distribution message to all of the distribution target nodes in the content distribution path.

Specifically, the content distribution message sent by the distribution controller includes the content distribution path and the identifier of the content to be distributed, and after the distribution controller distributes the content distribution path to each distribution target node, the distribution target node may The content distribution path determines from which node the content is received and to which node the content is sent. In addition, each distribution target node is also capable of receiving an identifier of the content to be distributed, and the distribution target node can determine which content should be specifically received or transmitted according to the identifier.

Preferably, the distribution controller can pre-open the HTTPS port on it, and the HTTPS port is used for secure transmission of instructions. The distribution controller can use the HTTPS port to send a content distribution message to the distribution destination node to improve the security of content distribution.

S207. The distribution target node establishes a content connection channel.

After each distribution target node receives the content distribution path sent by the distribution controller, a connection channel can be established with the other distribution target nodes according to the content distribution path. Specifically, the distribution target node can determine which nodes to connect with based on the content distribution path. For example, if there is a chain such as node A->node B->node C in the content distribution path, it means that when the distribution controller requests content distribution, the content is forwarded from node A to node B, and then forwarded by node B. To node C, for node B, a connection channel should be established with node A and node C.

Each distribution target node can open a TCP port in advance, and the TCP port can be used for secure transmission of content to be distributed. When a connection channel is specifically established, each distribution target node and its neighbor nodes can establish a connection channel through a TCP port.

S208. The distribution target node sends a request message for acquiring metadata of the content to be distributed to the head node.

On the basis of the completion of the establishment of the content connection channel between the distribution target nodes, each distribution target node can be from the content The metadata of the content to be distributed is obtained in the first node in the distribution path. As described above, the first node in the content distribution path is a node having the complete distribution content, and in addition, the metadata of the content to be distributed is also stored in the first node, and the metadata is obtained before the first node obtains the complete distribution content. Save to the first node. Since the content repository stores the metadata of the complete distribution content and the content to be distributed, the content warehouse can be used as the first node. FIG. 1 also takes the content warehouse as the first node as an example. However, it should be noted that the head node in a content distribution path is not limited to the content warehouse, and may be other nodes that have the complete distribution content and the metadata of the content to be distributed.

S209. The first node sends, to each distribution target node, a response message of the request message for acquiring metadata of the content to be distributed, where the response message includes metadata of the content to be distributed.

S210. The distribution target node reports the ready state to the distribution controller.

After the connection channel is established between the distribution target nodes, and the metadata of the content to be distributed has been acquired from the first node, each distribution target node may report a ready status message to the distribution controller to notify the distribution control. The device is ready to begin content distribution. Each distribution target node can send a ready-to-use body to the distribution controller through a different port, for example: an HTTPS port.

S211. The distribution controller sends a content receiving and forwarding indication to each distribution target node.

Specifically, after the distribution controller receives the ready status message of each distribution target node, the content receiving and forwarding indication may be sent to each distribution target node, and each distribution target node may start to perform only after receiving the indication. Content reception and forwarding.

It should be noted that, for a content distribution path, the distribution controller needs to receive the ready status information of each of the distribution target nodes before sending the content receiving and forwarding indication to each distribution target node.

Preferably, the distribution controller sends a content receiving and forwarding indication to the distribution target node through the HTTPS port.

S212. Perform content reception and forwarding between each distribution target node.

Specifically, each distribution target node receives the content to be distributed from the receiving channel according to the content connection channel established as described above, and forwards the content to be distributed through the forwarding channel.

The specific process of distributing the target node for content reception and forwarding will be explained in detail in the introduction of FIG. 14 below.

Preferably, the content distribution and forwarding are performed between the distribution target nodes through the TCP port.

S213. Each distribution target node acquires a distribution progress, a distribution log, a distribution result, and a status flow.

Specifically, each of the distribution target nodes can monitor the current receiving and forwarding progress in real time during the process of receiving and forwarding the content, and can also obtain the log information generated during the content receiving and forwarding process, and can also receive and forward, that is, the distribution is completed. After the distribution result is obtained, for example, which blocks are successfully received, whether the content forwarding is successful, or the like, the current state traffic of the distribution target node can be monitored in real time, wherein the state traffic can indicate the current state of the distribution target node and the current traffic, for example, Currently, it is in the normal state of connection or fault state, and the current value of the received and forwarded traffic.

The foregoing receiving and forwarding progress may refer to receiving the number of content blocks that the distribution target node has received, or the proportion of the content blocks that have been received occupying all the content blocks that need to be received, and the like.

S214. Each distribution target node reports the distribution progress, the distribution log, the distribution result, and the status traffic to the distribution controller.

Preferably, each distribution target node can periodically report the distribution progress, the distribution log, the distribution result, and the status traffic, and the distribution controller can save the information, analyze the information, and perform the corresponding operation according to the analysis result. For example, the distribution controller analyzes the state traffic of each node. If the state traffic of a node exceeds a certain threshold, the distribution controller can perform corresponding flow control; for example, the distribution controller can perform the distribution result. Analysis, if the distribution result is a failure, you need to re-distribute with the corresponding strategy.

In addition, the distribution controller can also send the above distribution progress, distribution log, distribution result, and status traffic to the terminal. Specifically, the terminal may query the foregoing information through an API provided by the distribution controller, so as to timely understand the content distribution and feed back the situation to the user.

Preferably, the distribution target node reports the progress, log, distribution result, and status traffic to the distribution controller through the HTTPS port.

Accordingly, the distribution controller also receives the progress of the distribution report, the log, the distribution result, and the status traffic through the HTTPS port.

Preferably, after the content is completed, the TCP connection can be disconnected between the distribution target nodes, the distribution target node reports the distribution to the distribution controller using the HTTPS port, and the distribution controller also receives the distribution target node report through the HTTPS port. Distribution completion information.

Table 2 provides an introduction to the ports that are open in the distribution controller, content store, and distribution target nodes and their functions.

Table 2

Using the https secure port for command transfer and distribution of content transfers ensures security during content distribution. Use the tcp port to transfer content more quickly, ensuring efficient content transfer.

In this embodiment, when content distribution is performed, the distribution controller generates a content distribution path according to the physical networking topology, the state traffic of each node in the system, and the distribution target node list, and further performs content distribution according to the content distribution path. First, the content distribution path is generated according to the physical networking topology information and the state traffic of each node in the system, so that it is possible to ensure that there is no disordered cross-connection between the distribution target nodes in the content distribution path, and that the content distribution path is ensured. The actual traffic of each distribution target node in the control range is within the controllable range. Secondly, the content distribution path is a unidirectional chain path, that is, the number of nodes connected to each node in the content distribution path is controlled by the distribution controller, so that the balance of the inbound and outbound bandwidth of each distribution target node can be achieved. In summary, it is possible to ensure that the selected content distribution path is the optimal content distribution path. In addition, in this embodiment, the distribution controller can generate the content distribution path according to the physical networking topology information and the state traffic of each node in the system each time the content distribution instruction is received, thereby ensuring that the content is performed each time. The content distribution paths generated at the time of distribution are the optimal distribution paths. Further, based on the above-described optimal content distribution path, the flow control in the content distribution process can be fully realized.

Further, in this embodiment, the content distribution path is generated by the distribution controller according to the distribution task situation, the content receiving node does not need to be dynamically added, the scheduling of the distribution controller is simple, and the content receiving node does not need to continuously query the available path during the distribution process. Therefore, it is also possible to ensure the efficiency of content distribution.

Further, in this embodiment, the distribution path is a unidirectional chain path combined with the actual physical topology networking structure, and the nodes are completely peer-to-peer, and the required file block information is not required to be queried. After the current node receives the distributed content, the current node can Forwarded to the next node immediately, and there is a fixed head node, so the integrity of the content distribution can also be guaranteed.

In another embodiment, in the above step S203, the distribution controller may also specify a maximum distribution rate of each node in the content distribution path when generating the content distribution path.

Specifically, the distribution controller determines the maximum distribution rate of each node according to the current state traffic of each node in the content distribution path, that is, the rate at which each node receives data, and if the current traffic of a node is low, The maximum distribution rate of the node is set to be larger. If the current traffic of a node has reached a certain threshold, the maximum distribution rate of the node needs to be set smaller.

Further, the distribution controller may carry the above maximum distribution rate in the content distribution message and send it to each node. The node constrains content distribution according to the maximum distribution rate when content distribution is performed.

In this embodiment, by setting a maximum distribution rate for each node in the content distribution path, it is ensured that the distribution traffic of each node is controllable, thereby further ensuring traffic control during content distribution.

The following is a detailed description of several specific implementation methods for the distribution controller to generate a content distribution path.

The first specific implementation method:

The distribution controller uses two distribution target nodes located on the same physical host in the physical networking as adjacent nodes in the content distribution path according to the physical networking topology and the distribution target node list.

Taking a cloud computing system as an example, FIG. 3 is an example of a central physical server of a cloud computing system, and FIG. 4 is an example of internal connection of a central physical server of a cloud computing system. As shown in FIG. 3 and FIG. 4, the central physical server may be A blade server is used, and the components of the blade server are as shown in Table 1:

Table 1

11	机箱 Chassis
22	计算节点，即物理主机Compute node
33	风扇模块Fan module
44	直通模块/交换模块(交换机)Pass-through module / switch module (switch)
55	管理模块Management module
66	电源模块Power module

Referring to FIG. 3, FIG. 4 and Table 1, the blade server in FIG. 3 has 16 computing nodes, that is, 16 physical hosts, each of which is configured with a 64-core CPU and 256 GB of memory, and each physical host can Virtually out 16-32 virtual machines. The NIC bandwidth of the physical host is determined, for example, 10 Gbps, and the bandwidth of the switching module in the entire chassis is also determined, for example, 10 Gbps. When the number of virtual machines corresponding to a physical host is large, the bandwidth that can be equally distributed to each virtual machine is very limited.

In this embodiment, specifically, the distribution server may use two distribution target nodes located on the same physical host in the physical networking as neighboring nodes in the content distribution path according to the physical networking topology when generating the content distribution path. . For example, suppose that the distribution target node received by the distribution server includes: node A, node B, node C, and node D, wherein node A and node B are in the same physical host 21, and node C and node D are in the same physical Within the host 22, the distribution controller can use the node B and the node A as neighboring nodes in the content distribution path, and the node D and the node C as the content distribution path. Adjacent nodes. In this way, the content receiving and forwarding between the node A and the node B are both in the physical host 21, and the network card bandwidth of the physical host 21 is not occupied. Similarly, the content receiving and forwarding between the node C and the node D are in the physical host. Within 22, the NIC bandwidth of the physical host 22 is not occupied. In this way, it is possible to generate a content distribution path that is ordered, has no intersection, and neighboring nodes are all located in the same physical host, thereby realizing flow control and high efficiency in content distribution.

The second specific implementation method:

According to the physical networking topology and the distribution target node list, two distribution target nodes located in the same chassis in the physical network are used as adjacent nodes in the content distribution path.

Specifically, referring to FIG. 3, FIG. 4 and Table 1, when there are no nodes located in the same physical host in the distribution target node received by the distribution controller, that is, each distribution target node does not satisfy the condition of being located on the same physical host. The distribution controller can then use two nodes located in the same chassis as neighbors in the content distribution path. In this way, the data interaction between the two nodes can be limited to the same chassis as much as possible, that is, the bandwidth of the chassis is not occupied as much as possible, thereby further ensuring the flow control and high efficiency in content distribution.

The third specific implementation method:

A node having a complete content to be distributed in the system is determined; according to the state traffic of each node in the system, the node with the lowest current traffic consumption among the nodes having the complete content to be distributed is used as the first node in the content distribution path.

As mentioned earlier, the head node needs to have complete distribution content and metadata for the content to be distributed. The first node can obtain the above-mentioned complete distribution content and metadata of the content to be distributed by the following method:

In the initial state, all the nodes in the system do not have the complete distribution content and the metadata of the content to be distributed, but only the content distribution stores the complete distribution content. When the user initiates the content distribution instruction, the content warehouse can be based on The complete distribution of content generates metadata for the content to be distributed. In this case, the first node in the content distribution path generated by the distribution controller is the content repository. After the distribution controller sends the content distribution path to each distribution destination node, each distribution destination node can be obtained from the content repository. Metadata of the content to be distributed, and establishing a connection with other nodes according to the content distribution path, and sending a ready state to the distribution controller, the distribution controller receives the ready state, and sends a notification to the head node (the first node is the content The warehouse, after receiving the notification, the first node forwards the content to its downstream node according to the connection relationship in the content distribution path. When a distribution target node receives all the content to be distributed, the distribution target node has the complete distribution content. At the same time, each distribution target node in the content distribution path can report the content information that has been successfully received and saved to the distribution controller in the process of receiving and forwarding the content, and the distribution controller can determine which nodes have been completed according to the information. Distribute content. Specifically, the distribution target node may report to the distribution controller a quantity of the content block that has been received, an identifier of each content block, a success indication of receiving, and the like according to the information, and the distribution controller determines, according to the information, whether the distribution target node is Already have a complete distribution.

When there are multiple nodes in the system with the complete distribution content, when the distribution controller generates the content distribution path for the complete distribution content, the node with the lowest current traffic consumption can be selected as the content distribution path in the current content distribution process. The first node.

In this embodiment, by selecting the node with the lowest current traffic consumption as the first node, the distribution speed of the content to be distributed can be ensured, thereby further improving the efficiency of content distribution.

The fourth specific implementation method:

If there are multiple nodes having complete content to be distributed, the generated content distribution path is multiple, wherein the number of generated content distribution paths is less than or equal to the number of nodes having complete content to be distributed, The first node in each content distribution path is one of the nodes having the complete content to be distributed.

Specifically, by default, the content distribution path generated by the distribution controller is a single chain path, and FIG. 5 is a single A schematic diagram of a chain path, as shown in FIG. 5, each node in a single chain path has only one upstream node and one downstream node, that is, the node can only receive data from the upstream node, and can only send data to the downstream node, thereby It can fully control the inbound and outbound bandwidth of each node, avoiding congestion caused by too many connections of a single node, or excessive consumption of CPU and memory, and affecting applications running on the node. The arrows in Figure 5 indicate the flow of data. With a small number of distribution target nodes, a single chained path ensures traffic control, efficiency, and integrity in content distribution. However, if the number of distribution target nodes is large, using a single chain path may result in a slower distribution due to the link being too long.

In this embodiment, when there are multiple nodes in the system that have complete content to be distributed, the nodes satisfy the condition as the first node, and the distribution controller can generate multiple content distribution paths according to the nodes. The first node in each content distribution path is one of these nodes with complete content to be distributed. The content distribution paths generated by the distribution controller are independent of each other, so parallel distribution between content distribution paths can be achieved.

Preferably, the distribution controller can set the nodes located in the same physical host or the same chassis in the same content distribution path, thereby improving content distribution efficiency and saving traffic consumption.

6 is a schematic diagram of parallel distribution of multiple content distribution paths. As shown in FIG. 6, the distribution controller sets a plurality of distribution target nodes in two content distribution paths by the above method, and the first in each content distribution path. The node is determined by the distribution controller according to information such as the number of content blocks that have been received reported by each distribution target node, the identifier of each content block, the success indication of receiving, and the like.

The method for parallel distribution of multiple content distribution paths in the embodiment can effectively reduce the influence of the delay of the last node distribution on the link due to the long link length, and can reduce the complexity caused by the length of the control chain path. degree.

The fifth specific implementation method:

Set the number of preset nodes in the content distribution path; if the number of nodes in the content distribution path is greater than the preset number of nodes, select the distribution target node with the lowest current traffic consumption as the branch node from the content distribution path, and according to the score The fork node splits the content distribution path into a first content distribution path and a second content distribution path.

The first content distribution path and the second content distribution path are the same as the first node, and the first content distribution path and the second content distribution path are demarcated by the foregoing bifurcation node.

Specifically, the distribution controller determines the node with the lowest current traffic consumption according to the state traffic reported by the node, and uses the distribution target node with the lowest current traffic consumption as the fork node.

Specifically, the distribution controller presets the preset number of nodes according to the physical networking topology and the node size of the system, for example, setting the number of nodes to 200. When the number of distribution target nodes is less than or equal to the preset number of nodes, the distribution controller may Generating a single chained distribution path, using this single chained distribution path for content distribution ensures the efficiency of content distribution. If the number of the distribution target nodes is greater than the number of the preset nodes, if the single-chain distribution path is used again, the delay of the last node in the path is too large. Therefore, in this embodiment, the method may be firstly performed according to the foregoing method. Establishing a single chain distribution path, and then selecting a node with the lowest current traffic as the branch node from the single chain distribution path, and after the fork node, generating two distribution paths, that is, after the branch node , the original distribution path is changed into two paths, thus reducing the length of the distribution path and implementing parallel content distribution. Since the number of nodes of the forked node is two, the amount of traffic required is relatively large. Therefore, selecting the node with the lowest current traffic as the fork node can ensure that the fork node has sufficient traffic for the fork node. Content distribution.

FIG. 7 is a schematic diagram of the content distribution path bifurcation. As shown in FIG. 7, the first node after the first node is used as a fork node, and two parallel distribution paths are formed from the node, thereby reducing distribution. The length of the path ensures the speed and efficiency of content distribution.

It should be noted that, in the foregoing fourth specific implementation manner, whether the distribution target node is greater than a preset may be first determined. The number of nodes, if greater than the above, implements the fourth embodiment described above.

The sixth specific implementation method:

This embodiment relates to a method of processing a faulty node in an existing content distribution path.

When the distribution target node on the content distribution link performs content distribution, data may not be normally received or forwarded for some reason, in which case the distribution controller first determines the faulty node in the content distribution path.

The method for the distribution controller to determine the faulty node in the content distribution path includes the following three types:

After the distribution controller sends the content distribution message to the first distribution target node, if the response message for the content distribution message sent by the first distribution target node is not received, determining that the first distribution target node is in the content distribution path A faulty node, wherein the first distribution target node is a distribution target node of any one of the content distribution paths.

The specific process of the distribution controller transmitting the content distribution message to the first distribution target node is as described in the foregoing step S206. If the distribution target node is operating normally, after the content distribution message sent by the distribution controller is received, the distribution target node distributes to the distribution node. Control returns a response message. If the distribution controller does not receive the response message, the distribution controller can determine the distribution target node failure.

2. Determine the faulty node in the content distribution path according to the status traffic reported by the at least one distribution target node and the distribution log.

After the distribution target node starts content distribution, the distribution target node can report its status traffic and distribution logs to the distribution controller according to a certain period, and the distribution controller receives the status traffic and distributes the log, and then judges if the status of the target node is distributed. If the traffic has exceeded the maximum threshold, or there is a record of the distribution exception in the distribution log, the distribution controller can determine that the distribution target node is down.

3. If the distribution controller receives the distribution failure message sent by the second distribution target node, determining that the downstream node of the second distribution target node is a faulty node in the content distribution path, wherein the second distribution target node is a content distribution path. Any one of the distribution target nodes except the last distribution target node.

The distribution target node may record the content block that has been distributed in the process of content distribution, and if there is a case where the content block distribution fails, the distribution target node sends a distribution failure message to the distribution controller. After receiving the message, the distribution controller may determine that the downstream node of the distribution target node cannot receive the distribution content normally, and determine the downstream node of the distribution target node as the fault node.

After the distribution controller determines the faulty node in the content distribution path, the distribution controller may delete the faulty node from the original content distribution path and use the downstream node of the faulty node as the downstream node of the upstream node of the faulty node. Preferably, the distribution controller separately sends a path update indication to the upstream node and the downstream node of the faulty node to indicate that the upstream node of the faulty node updates its saved content distribution path, that is, updates its downstream node to the downstream node of the faulty node. Indicates that the downstream node of the faulty node updates its saved content distribution path, that is, updates its upstream node to the upstream node of the faulty node. It should be noted that after the above update is completed, a new connection channel needs to be established between the upstream node and the downstream node of the faulty node to ensure normal distribution of subsequent content.

8 is a schematic diagram of deleting a faulty node from a content distribution path. As shown in FIG. 8, the second node in the content distribution path is a faulty node, and the faulty node is deleted from the content distribution path, and the faulty node is before and The nodes that follow are connected.

The seventh specific embodiment:

This embodiment relates to a processing method after a failed node is recovered.

After the faulty node is deleted from the content distribution path, the faulty node can continue to periodically report its state traffic to the distribution controller. When the distribution controller knows that the faulty node's state has returned to the normal state, it joins again. In the content distribution path.

Specifically, the distribution controller acquires the recovery node in the content distribution path by acquiring the state traffic. The number of the recovery nodes may be one or multiple, and the recovery node is a node deleted from the content distribution path due to the failure. After the distribution controller acquires the recovery node, the distribution controller may add the recovery node to the content distribution path to cause the recovery node to obtain the distribution content from the upstream node of the recovery node, wherein the recovery node requests the slave node from the upstream node. Get the distribution. Specifically, FIG. 9 is a schematic diagram of rejoining the recovery node to the content distribution path, and the dotted line in FIG. 9 indicates that the node A and the node B are in the content distribution path in the normal state, and the node A and the node B are from the content distribution path due to the failure. In the middle deletion, the upstream node of node A is connected with the downstream node, and the upstream node of node B is connected with the downstream node. When the state of node A and node B returns to the normal state, the content distribution path does not return to the original state. It is only allowed that node A obtains the distributed content from its upstream node by means of an active request, and only allows the node B to obtain the distributed content from its upstream node by means of an active request. It should be noted that the upstream node at this time is not necessarily the upstream node of the restoration node in the original content distribution path, but the node that is closest to the recovery node in the physical topology and is in the current content distribution path.

The eighth specific embodiment:

For some systems, such as cloud computing systems, there may be situations where the same distribution is distributed to multiple subnets. 10 is a schematic diagram of content distribution to multiple subnets. As shown in FIG. 10, in the case where content is distributed to multiple subnets, the distribution controller may divide the content distribution path into subnets based on the first node. Multiple sub-paths, no isolation between multiple sub-paths. This effectively controls the bandwidth usage of each subnet and avoids traffic interaction between subnets.

Further, if the length of the sub-path within a subnet is too long, the sub-path can be split according to the method described in the foregoing embodiment to ensure the speed and efficiency of content distribution.

Based on the above embodiment, the present embodiment relates to a specific method for generating metadata of the content to be distributed in the content repository of step 105 in FIG. 2 described above.

As described in step S205 in FIG. 2, after the content repository receives the creation metadata indication, the content to be distributed stored in the content repository is determined according to the identifier of the content to be distributed carried therein, and the content to be distributed is generated. Metadata. Before generating the metadata of the content to be distributed, the content repository first determines whether the metadata of the content to be distributed has been generated, and if it has been generated, the generation operation is no longer performed, but the existing metadata can be directly used.

At the same time, the content repository may determine whether the content to be distributed exists according to the identifier of the content to be distributed in the created metadata indication. If not, the content warehouse may return a prompt to the terminal that the content to be distributed does not exist. The terminal is caused to re-check the identity of the content to be distributed that is sent, or to re-publish the content to be distributed to the content repository.

The identifier of the content to be distributed is allocated by the content warehouse, and is different from the identifier corresponding to other content to be distributed in the entire system.

If the content repository confirms that the content to be distributed exists and the metadata of the content to be distributed does not yet exist, it is necessary to generate metadata of the content to be distributed. In the subsequent content distribution process, if the metadata is not saved in the nodes in the content distribution path, the metadata of the content to be distributed can be obtained from the content repository.

Specifically, FIG. 11 is a schematic diagram of a process for generating metadata according to Embodiment 2 of the present invention. As shown in FIG. 11, a method for generating metadata of a content to be distributed by a content warehouse is:

S1101: Perform partial division and block division on the content to be distributed, and obtain a partial quantity, a partial size, a block size, and a block index.

The size of the content to be distributed is generally large. For example, a general multimedia file that needs to be distributed can reach more than 1G. If the entire content to be distributed is sent directly in the content distribution process, once the data transmission has a problem, it needs to be heavy. Newly sent the entire file to be distributed, resulting in inefficient content distribution. Therefore, in this embodiment, the content to be distributed is hierarchically divided into blocks. In the content distribution process, the minimum unit of data transmission is blocking, and even if there is a problem in transmission, only the corresponding segment needs to be redistributed. Can greatly improve the efficiency of content distribution.

The content warehouse can divide the content to be distributed into two layers. The first layer divides the entire file to be distributed into a plurality of parts, and the second layer divides each part into a plurality of blocks. FIG. 12 is a schematic diagram of hierarchical division of content to be distributed. As shown in FIG. 12, the content repository may divide the content to be distributed into multiple equal-sized parts, and then divide each part into blocks of equal size, and the block is content distribution. The minimum transmission unit.

The user can pre-configure the size of the part and the size of the block in the content repository. When the content warehouse is hierarchically divided, the partial partitioning and block partitioning are performed according to the pre-configured partial size and block size. It should be noted that, according to requirements, the content warehouse can also divide the content to be distributed into more layers, wherein the lowest layer is the smallest unit of data transmission.

In this step, when the content warehouse performs partial partitioning and block partitioning, the partial quantity and the block number are calculated according to the total size, partial size, and block size of the content to be distributed, and the total size of the content to be distributed may be specifically divided by Partially sized to get a partial quantity, then divide the partial size by the block size to get the number of blocks. Moreover, after the content warehouse divides the blocks, each block can also be assigned a unique index, that is, a block index, for identifying each block when subsequent content distribution. After the process, the content warehouse can get the partial quantity, partial size, block size and block index.

When assigning a block index to a block, preferably, the index value may be determined according to the number of the portion of the block corresponding to the block and the order of the block in the portion. For example, suppose a block M, the part corresponding to the block M is the first part of the content to be distributed, that is, the corresponding part number is 1, and the block is the first block in the part 1, that is, the block If the number is 1, the index value of the block can be set to P1B1, where "P" refers to the part and "B" refers to the block.

Since the partial size and the block size are fixed, the content corresponding to the block can be immediately indexed according to the block index. Therefore, the allocation of the block index according to the number can facilitate reading and writing of the distributed content by offset. Can improve the efficiency of reading and writing of distributed content.

S1102: Perform digital signature processing on the content to be distributed, and obtain a digital signature.

The content warehouse can use a specific algorithm, such as (Pretty Good Privacy, PGP for short) to digitally sign the distributed content, obtain a digital signature, and use the same algorithm pair after the distribution target node obtains the complete content to be distributed. The received content is checked for signature to confirm the security of the content to be distributed.

S1103: Perform hash check on each block that is divided, and obtain a block hash value of each block.

In addition to digitally signing the entire content to be distributed, the content repository can also hash each block that is partitioned. Since the block is the smallest unit of data transmission, the content warehouse performs a hash check on each block in advance and writes it into the metadata of the content to be distributed. After the distribution target node receives a content block, it can It is determined whether the received block has not been tampered by comparing the hash check results to ensure the security of each block. When hashing a block, you can use the Secure Hash Algorithm (SHA1) algorithm.

S1104: Write a partial number, a partial size, a block size, a block index, a digital signature, and a block hash value of each block into the metadata of the content to be distributed.

After the partial number, partial size, block size, block index, digital signature, and block hash value of each block are acquired through the foregoing steps (S1101-1104), the content repository can write it into the metadata of the distributed content.

13 is an exemplary diagram of metadata of content to be distributed. As shown in FIG. 13, the first part of the metadata of the content to be distributed includes the name of the content to be distributed and the digital signature, and the second part includes the part count and the partial size. (partsize) and block size (blocksize), the third part includes each block index and block hash value (blockhash) of each block. Among them, in the third part, the block hash value of each block is after each block index, so After the target node reads the block index, it can directly read the block hash value corresponding to the block without querying again, thereby improving the reading efficiency of the metadata.

Based on the foregoing embodiment, the embodiment relates to the specific method for distributing the target node for content distribution in step S213 in FIG. 2.

As described in steps S206-S213 in the foregoing first embodiment, after the distribution target node receives the content distribution message, performing the operation of establishing a connection, acquiring metadata of the content to be distributed, reporting the ready state, and then distributing according to the metadata and the content. The path performs content distribution processing, that is, content reception and forwarding.

FIG. 14 is a schematic diagram of a content receiving and forwarding process according to Embodiment 3 of a content distribution method according to an embodiment of the present invention. As shown in FIG. 14, the process of distributing a target node for content distribution is:

S1401. Receive a block sent by an upstream node.

S1402: Perform a hash check on the block according to the block index and the block hash value in the metadata of the content to be distributed, and obtain the verification result.

S1403. If the verification result is that the verification is successful, the block is saved, and the block is sent to the downstream node in the content distribution path according to the specified distribution rate.

The distribution rate refers to the rate of reception, because the number of ingress and egress nodes is fixed, so the forwarding rate is also controlled.

Specifically, the distribution target node performs hash check on the received block to obtain a check value, and then reads the block hash value from the metadata according to the block index of the received block, if the distribution target node calculates If the check value is consistent with the block hash value in the metadata, the received block is considered to be the correct block and can be forwarded to the next distribution target node.

As described above, the content to be distributed is in the block as the minimum transmission unit, and therefore, each time the distribution target node receives a block, the above-described steps S1401 - S1403 are performed.

In this embodiment, after the distribution target node receives a block, the block can be verified by using a hash check, and if the verification is successful, it is forwarded to the next distribution target node, thereby ensuring the forwarded block. safety.

It should be noted that, for the first node in the content distribution path, after receiving the content distribution indication sent by the distribution controller, the content can be directly forwarded to the downstream node, and therefore, the above steps S1401-S1403 need not be performed.

In another embodiment, when the block is forwarded to the downstream node in step S1403, the distribution target node may determine whether the transmission is successful. If the transmission is not successful, the retransmission is performed. If the number of retransmission failures reaches a preset number, the The distribution controller sends a distribution failure message to cause the distribution controller to re-establish the content distribution path according to the state of the downstream node.

In another embodiment, after the step S1403, the distribution target node further needs to perform the following operations:

The distribution target node determines whether all the blocks corresponding to the content to be distributed are received, and if so, the distribution target node performs a digital signature check on the content to be distributed according to the digital signature in the metadata of the content to be distributed, and reports the result of the digital signature check to the distribution. Controller.

Specifically, as described above, the digital signature check is used to check the entire content to be distributed, and therefore, after the distribution target node confirms that the blocks of the content to be distributed have all been received, the corresponding algorithm can be used to receive the entire The content is digitally checked. If the digital signature result is consistent with the metadata of the content to be distributed, the check passes, otherwise the check fails. The distribution target node reports the digital signature result to the distribution controller, and the distribution controller performs retransmission and the like according to the digital signature result.

The distribution target may determine whether all blocks of the content to be distributed have been received according to the saved metadata of the content to be distributed. Specifically, each block index of the content to be distributed is recorded in the metadata, and the distribution control determines whether the block index of all the received content blocks completely corresponds to the block index in the metadata, and if yes, determines that the to-be-score has been received. Send all blocks of content.

The following embodiments relate to the scenario in which the present invention is applied to a cloud computing system in which a plurality of tenants exist.

For a cloud computing system, where there are multiple tenants, the cloud computing system requires that the resources of multiple tenants are isolated from each other. Therefore, when the embodiment of the present invention is applied to a cloud computing system, it needs to be based on the characteristics of the tenant. Content distribution. In multi-tenant technology, a tenant is a customer who uses a system or computer computing resource, but in multi-tenant technology, a tenant contains all data that can be identified as a specified user in the system, including account and statistics (accounting data) ), the various data built by the user in the system, and the customized application environment of the user itself are all in the scope of the tenant, and the tenant uses the application system developed or built by the supplier or For computing resources, etc., the application system designed by the vendor will accommodate more than one user in the same environment. In order to make the environment capabilities of multiple users the same application and computing environment, the application and computing environment It must be specially designed, in addition to allowing the system platform to allow multiple copies of the same application to run at the same time.

Specifically, when the user or the application invokes the API provided by the distribution controller to deliver the content distribution indication to the distribution controller, the content distribution indication may be carried in addition to the identifier of the to-be-distributed content and the distribution target node. Tenant information. After receiving the content distribution instruction, the distribution controller may integrate the unified user management system of the third party to authenticate the tenant and query the resource list accessible by the tenant. Further, when the distribution controller constructs the content distribution path, the node in the content distribution path is a distribution target node accessible by the tenant, and the node in the content distribution path cannot be a distribution target node inaccessible to the tenant. Different tenants build different content distribution paths, and different content distribution paths cannot access each other during the distribution process, thereby effectively achieving multi-tenancy isolation.

Figure 15 is a block diagram of a first embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in Figure 15, the apparatus includes:

The receiving module 1501 is configured to receive a content distribution indication, where the content distribution indication includes an identifier of the content to be distributed and a distribution target node list, where the distribution target node list includes an identifier of the at least one distribution target node.

The processing module 1502 is configured to generate a content distribution path according to the physical networking topology, the status traffic of each node in the system, and the distribution target node list, where the content distribution path is a single page formed by the first node and the at least one distribution target node. The chained distribution path, the first node is the node in the system with complete content to be distributed.

The sending module 1503 is configured to send a content distribution message to the at least one distribution target node, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, so that the at least one distribution target node obtains according to the content distribution path and the identifier of the content to be distributed. And distribute the content to be distributed.

The device is used to implement the foregoing method embodiments, and the implementation principle and technical effects are similar, and details are not described herein again.

In another embodiment, the content distribution message further includes a specified distribution rate of the at least one distribution target node, where the specified distribution rate is determined according to the state traffic of the at least one distribution target node when the content distribution path is generated.

In another embodiment, the receiving module 1501 is further configured to receive a ready state reported by the one distribution target node.

The sending module 1503 is further configured to send a content receiving and forwarding indication to the at least one distribution target node according to the ready state.

Figure 16 is a block diagram of a second embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in Figure 16, the processing module includes:

The first determining unit 15021 is configured to use two distribution target nodes that are located on the same physical host or the same chassis in the physical networking as the neighboring nodes in the content distribution path according to the physical networking topology and the distribution target node list.

The second determining unit 15022 is configured to determine a node in the system that has complete content to be distributed.

The third determining unit 15023 is configured to use, as the first node in the content distribution path, the node with the lowest current traffic consumption among the nodes having the complete content to be distributed according to the state traffic of each node in the system.

In another embodiment, the processing module 1502 is further configured to:

Determining multiple nodes in the system with complete content to be distributed; and,

Generating a plurality of content distribution paths, wherein the number of generated content distribution paths is less than or equal to the number of nodes having complete content to be distributed, and the first node in each content distribution path is complete with content to be distributed One of the nodes in the node.

In another embodiment, the processing module 1502 is further configured to:

Set the number of preset nodes in the content distribution path; and,

If the number of nodes in the content distribution path is greater than the preset number of nodes, selecting a fork node from the content distribution path, and splitting the content distribution path into the first content distribution path and the second content distribution path according to the fork node;

The first content distribution path and the second content distribution path are the same as the first node, the first content distribution path and the second content distribution path are demarcated by the fork node, and the fork node is the node with the lowest current traffic consumption in the content distribution path. .

In another embodiment, the processing module 1502 is further configured to:

Determining the faulty node in the content distribution path; and,

The faulty node is deleted from the content distribution path, and the downstream node of the faulty node is regarded as the downstream node of the upstream node of the faulty node.

FIG. 17 is a block diagram of a third embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 17, the processing module 1502 further includes:

The fourth determining unit 15024 is configured to: when the response message for the content distribution message sent by the first distribution target node is not received, determine that the first distribution target node is a faulty node in the content distribution path, where the first distribution target The node distributes the target node for any of the content distribution paths.

FIG. 18 is a block diagram of a fourth embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 18, the processing module 1502 further includes:

The fifth determining unit 15025 is configured to determine a faulty node in the content distribution path according to the state traffic reported by the at least one distribution target node and the distribution log.

FIG. 19 is a block diagram of a fifth embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 19, the processing module 1502 further includes:

The sixth determining unit 15026 is configured to: when receiving the distribution failure message sent by the second distribution target node, determine that the downstream node of the second distribution target node is a faulty node in the content distribution path, where the second distribution target node is Any one of the distribution target nodes except the last distribution destination node in the content distribution path.

In another embodiment, the processing module 1502 is further configured to:

The recovery node is added to the content distribution path such that the recovery node acquires the distribution content from the upstream node of the recovery node, wherein the recovery node acquires the distribution content from the upstream node of the recovery node by the active request mode.

In another embodiment, the receiving module 1501 is further configured to receive a distribution progress, a distribution log, a distribution result, and a status flow periodically reported by the at least one distribution target node.

The sending module 1503 is further configured to send a distribution progress, a distribution log, a distribution result, and the status traffic to the terminal.

FIG. 20 is a block diagram of a first embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 20, the apparatus includes:

The receiving module 2001 is configured to receive a content distribution message, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, where the content distribution path is configured by the distribution controller according to a physical networking topology, status traffic of each node in the system, and The distribution target node list is generated by a unidirectional chain distribution path composed of a first node and at least one distribution target node, and the first node is a node having complete content to be distributed in the system.

The processing module 2002 is configured to obtain metadata of the content to be distributed from the first node according to the identifier of the content to be distributed, where the metadata of the content to be distributed includes a digital signature, a block index, and a block hash value of the content to be distributed.

The processing module 2002 is further configured to perform content distribution processing according to the content distribution path and the metadata of the content to be distributed.

In another embodiment, the content distribution message further includes a specified distribution rate of the distribution target node.

The processing module 2002 is specifically configured to perform content distribution processing according to the specified distribution rate according to the content distribution path and the metadata of the content to be distributed.

Figure 21 is a block diagram of a second embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in Figure 21, the apparatus further includes:

The sending module 2003 is configured to report the ready state to the distribution controller.

The receiving module 2001 is further configured to receive a content receiving and forwarding indication sent by the distribution controller.

FIG. 22 is a block diagram of a third embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 22, the processing module 2002 includes:

The receiving unit 20021 is configured to receive a block sent by the upstream node.

The checking unit 20022 is configured to perform hash check on the block according to the block index and the block hash value in the metadata of the content to be distributed, and obtain the verification result.

The sending unit 20023 is configured to save the block when the verification result is that the verification is successful, and send the block to the downstream node in the content distribution path according to the specified distribution rate.

In another embodiment, the sending unit 20023 is specifically configured to:

If the transmission fails, resend; and,

In another embodiment, the processing module 2002 is further configured to perform a digital signature check on the content to be distributed according to the digital signature when receiving all the blocks corresponding to the content to be distributed, and report the result of the digital signature check to the distribution controller.

In another embodiment, the processing module 2002 is further configured to obtain a distribution progress, a distribution log, a distribution result, and a status flow.

The sending module 2003 is further configured to report the distribution progress, the distribution log, the distribution result, and the status traffic to the distribution controller.

FIG. 23 is a block diagram of a first embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 23, the apparatus includes:

The receiving module 2301 is configured to receive a creation metadata indication, where the creation metadata indication includes an identifier of the content to be distributed;

The processing module 2302 is configured to generate metadata of the content to be distributed according to the identifier of the content to be distributed.

FIG. 24 is a block diagram of a second embodiment of a content distribution apparatus according to an embodiment of the present invention. As shown in FIG. 24, the processing module 2302 includes:

The dividing unit 23021 is configured to perform partial division and block division on the to-be-distributed content after determining that the content to be distributed exists according to the content identifier to be distributed, and obtain a partial quantity, a partial size, a block size, and a block index.

The processing unit 23022 is configured to perform digital signature processing on the content to be distributed to obtain a digital signature.

The checking unit 23023 is configured to perform hash check on each of the divided blocks to obtain a block hash value of each block.

The writing unit 23024 is configured to write the partial quantity, the partial size, the block size, the block index, the digital signature, and the block hash value of each block into the metadata of the content to be distributed.

FIG. 25 is a physical block diagram of Embodiment 1 of a distribution controller according to an embodiment of the present invention. As shown in FIG. 25, the distribution controller includes:

A memory 2501 and a processor 2502.

The memory 2501 is used to store program instructions, and the processor 2502 is used to call program instructions in the memory 2501 to perform the following methods:

Generating a content distribution path according to a physical networking topology, a status traffic of each node in the system, and the distribution target node list, wherein the content distribution path is a one-way composed of a first node and the at least one distribution target node a chain distribution path, the head node being a node in the system having the complete content to be distributed;

Further, the content distribution message further includes a specified distribution rate of the at least one distribution target node, where the specified distribution rate is determined according to state traffic of the at least one distribution target node when the content distribution path is generated. .

Further, the processor 2502 is further configured to:

Receiving a ready state reported by the at least one distribution target node;

And transmitting, according to the ready state, a content receiving and forwarding indication to the at least one distribution target node.

Further, the processor 2502 is further configured to:

Determining a node in the system that has the complete content to be distributed;

The node having the lowest current traffic consumption among the nodes having the complete content to be distributed is used as the first node in the content distribution path according to the state traffic of each node in the system.

Further, the processor 2502 is further configured to:

Determining a plurality of nodes in the system having the complete content to be distributed;

Further, the processor 2502 is further configured to:

Setting a preset number of nodes in the content distribution path;

Further, the processor 2502 is further configured to:

Determining a faulty node in the content distribution path;

Further, the processor 2502 is further configured to:

FIG. 26 is a physical block diagram of Embodiment 1 of a computing node according to an embodiment of the present invention. As shown in FIG. 26, the computing node includes:

Memory 2601 and processor 2602.

The memory 2601 is configured to store program instructions, and the processor 2602 is configured to call program instructions in the memory 2601 to perform the following methods:

Receiving a content distribution message, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, wherein the content distribution path is distributed by the distribution controller according to a physical networking topology, status traffic of each node in the system, and a distribution target node The content distribution path is a unidirectional chain distribution path formed by the head node and the at least one distribution target node, where the head node is a node having a complete content to be distributed in the system;

The content distribution processing is performed according to the content distribution path and the metadata of the content to be distributed.

Further, the content distribution message further includes a specified distribution rate of the distribution target node; the processor 2602 is further configured to:

The content distribution processing is performed according to the content distribution path and the metadata of the content to be distributed according to the specified distribution rate.

Further, the processor 2602 is further configured to:

Reporting the ready status to the distribution controller;

Further, the processor 2602 is further configured to:

Receiving a block sent by an upstream node;

Further, the processor 2602 is further configured to:

If the transmission fails, resend it;

Further, the processor 2602 is further configured to:

FIG. 27 is a physical block diagram of Embodiment 1 of a content repository according to an embodiment of the present invention. As shown in FIG. 27, the content repository includes:

A memory 2701 and a processor 2702.

The memory 2701 is used to store program instructions, and the processor 2702 is used to call program instructions in the memory 2701 to perform the following methods:

Further, the processor 2702 is further configured to:

Hashing each block of the partition to obtain a block hash value of each block;

Write the partial number, the size of each part, the block size, the block index, the digital signature, and the block hash value of each block In the metadata of the content to be distributed.

In addition, an embodiment of the present invention further provides a content distribution system, including the foregoing distribution controller, the foregoing computing node, and the content repository.

One of ordinary skill in the art will appreciate that all or part of the steps to implement the various method embodiments described above may be accomplished by hardware associated with the program instructions. The aforementioned program can be stored in a computer readable storage medium. The program, when executed, performs the steps including the foregoing method embodiments; and the foregoing storage medium includes various media that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will understand that The technical solutions described in the foregoing embodiments may be modified, or some or all of the technical features may be equivalently replaced; and the modifications or substitutions do not deviate from the technical solutions of the embodiments of the present invention. range.

Claims

A content distribution method for a cloud computing system, comprising:

Receiving a content distribution indication, where the content distribution indication includes an identifier of the content to be distributed and a distribution target node list, where the distribution target node list includes an identifier of the at least one distribution target node;

Generating a content distribution path according to a physical networking topology of the cloud computing system, state traffic of each node, and the distribution target node list;

And transmitting, to the at least one distribution target node, a content distribution message, where the content distribution message includes the content distribution path and an identifier of the content to be distributed.
The method according to claim 1, wherein said content distribution message further comprises a specified distribution rate of said at least one distribution target node.
The method of claim 1 further comprising:

Receiving a ready state reported by the at least one distribution target node;

And transmitting, according to the ready state, a content receiving and forwarding indication to the at least one distribution target node.
The method according to claim 1, wherein the generating a content distribution path according to the physical networking topology of the cloud computing system, the state traffic of each node in the cloud computing system, and the distribution target node list comprises:

According to the physical networking topology of the cloud computing system and the distribution target node list, two distribution target nodes located on the same physical host or the same chassis in the physical networking are used as adjacent nodes in the content distribution path;

Determining a node in the cloud computing system that has the complete content to be distributed;

According to the state traffic of each node in the cloud computing system, the node with the lowest current traffic consumption among the nodes having the complete content to be distributed is used as the first node in the content distribution path.
The method of claim 1 further comprising:

According to the physical networking topology and the distribution target node list, two distribution target nodes located on the same physical host or the same chassis in the physical networking are used as adjacent nodes in the content distribution path;

Determining a plurality of nodes in the cloud computing system having the complete content to be distributed;

A plurality of content distribution paths are generated, wherein the number of generated content distribution paths is less than or equal to the number of nodes having the complete content to be distributed.
The method of claim 1 further comprising:

Setting a preset number of nodes in the content distribution path;

If the number of nodes in the content distribution path is greater than the preset number of nodes, selecting a fork node from the content distribution path, and splitting the content distribution path into the first content according to the fork node a distribution path and a second content distribution path;

The first content distribution path and the first node of the second content distribution path are the same, and the first content distribution path and the second content distribution path are demarcated by the bifurcation node, and the bifurcation The node is the node with the lowest traffic consumption in the content distribution path.
The method of claim 1 further comprising:

Determining a faulty node in the content distribution path;

And deleting the faulty node from the content distribution path, and using a downstream node of the faulty node as a downstream node of the upstream node of the faulty node.
The method according to claim 7, wherein said determining said faulty node in said content distribution path, include:

If the response message to the content distribution message sent by the first distribution target node is not received, determining that the first distribution target node is a faulty node in the content distribution path, where the first distribution target The node distributes the target node to any one of the content distribution paths;

Determining a faulty node in the content distribution path according to status traffic reported by the at least one distribution target node and a distribution log; or

If the distribution failure message sent by the second distribution target node is received, determining that the downstream node of the second distribution target node is one of the content distribution paths, wherein the second distribution target node is the Any one of the distribution target nodes except the last distribution destination node in the content distribution path.
The method of claim 7 further comprising:

Acquiring a recovery node in the content distribution path, the recovery node being a node deleted from the content distribution path due to a failure;

Adding the recovery node to the content distribution path, so that the recovery node acquires distribution content from an upstream node of the recovery node, wherein the recovery node is from an upstream node of the recovery node by an active request mode Get the distribution.
A content distribution method for a cloud computing system, comprising:

Receiving a content distribution message, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, wherein the content distribution path is configured by the distribution controller according to a physical networking topology of the cloud computing system, and each node in the cloud computing system The status flow is generated by the distribution target node list, and the content distribution path is a unidirectional chain distribution path formed by the first node and the at least one distribution target node, where the first node is complete in the system. a node that distributes content;

Obtaining, according to the identifier of the content to be distributed, the metadata of the content to be distributed from the first node, where the metadata of the content to be distributed includes a digital signature, a block index, and a block hash value of the content to be distributed;

The content distribution processing is performed according to the content distribution path and the metadata of the content to be distributed.
The method according to claim 10, wherein the content distribution message further comprises a specified distribution rate of the distribution target node;

The content distribution processing according to the content distribution path and the metadata of the content to be distributed specifically includes:

The content distribution processing is performed according to the content distribution path and the metadata of the content to be distributed according to the specified distribution rate.
The method of claim 11 further comprising:

Reporting the ready status to the distribution controller;

Receiving a content receiving and forwarding indication sent by the distribution controller.
The method according to claim 10, wherein the performing content distribution processing according to the to-be-content distribution path and the metadata of the content to be distributed includes:

Receiving a block sent by an upstream node;

And performing a hash check on the block according to the block index and the block hash value in the metadata of the content to be distributed, and acquiring a verification result;

If the verification result is that the verification is successful, the block is saved, and the block is sent to a downstream node in the content distribution path according to the specified distribution rate.
The method according to claim 13, wherein the sending the block to the downstream node in the content distribution path according to the specified distribution rate comprises:

If the transmission fails, resend it;

If the number of retransmission failures reaches a preset number of times, a distribution failure message is sent to the distribution controller to cause the distribution controller to re-establish the content distribution path according to the state of the downstream node.
The method according to claim 13 or 14, further comprising:

If all the blocks corresponding to the content to be distributed are received, the digital signature is checked according to the digital signature, and the result of the digital signature check is reported to the distribution controller.
A content distribution device, comprising:

a receiving module, configured to receive a content distribution indication, where the content distribution indication includes an identifier of the content to be distributed and a distribution target node list, where the distribution target node list includes an identifier of the at least one distribution target node;

a processing module, configured to generate a content distribution path according to a physical networking topology of the cloud computing system, a state traffic of each node in the cloud computing system, and the distribution target node list;

And a sending module, configured to send a content distribution message to the at least one distribution target node, where the content distribution message includes the content distribution path and an identifier of the content to be distributed.
The apparatus according to claim 16, wherein said content distribution message further comprises a specified distribution rate of said at least one distribution target node.
The device of claim 16 wherein:

The receiving module is further configured to receive a ready state reported by the at least one distribution target node;

The sending module is further configured to send a content receiving and forwarding indication to the at least one distribution target node according to the ready state.
The device according to claim 16, wherein the processing module comprises:

a first determining unit, configured to use, as the content distribution path, two distribution target nodes located on the same physical host or the same chassis in the physical networking according to the physical networking topology of the cloud computing system and the distribution target node list Adjacent nodes in ;

a second determining unit, configured to determine a node in the cloud computing system that has the complete content to be distributed;

And a third determining unit, configured to use, as the first node in the content distribution path, a node with the lowest current traffic consumption among the nodes having the complete content to be distributed according to the state traffic of each node in the cloud computing system.
The device according to claim 16, wherein the processing module is further configured to:

According to the physical networking topology and the distribution target node list, two distribution target nodes located on the same physical host or the same chassis in the physical networking are used as adjacent nodes in the content distribution path;

Determining a plurality of nodes in the system having the complete content to be distributed; and,

Generating a plurality of content distribution paths, wherein the number of generated content distribution paths is less than or equal to the number of nodes having the complete content to be distributed, and the first node in each of the content distribution paths is One of the nodes having the complete content to be distributed.
The device according to claim 16, wherein the processing module is further configured to:

Setting a preset number of nodes in the content distribution path; and,

If the number of nodes in the content distribution path is greater than the preset number of nodes, selecting a fork node from the content distribution path, and splitting the content distribution path into the first content according to the fork node a distribution path and a second content distribution path;

The first content distribution path and the first node of the second content distribution path are the same, and the first content distribution path and the second content distribution path are demarcated by the bifurcation node, and the bifurcation The node is the node with the lowest traffic consumption in the content distribution path.
The device according to claim 16, wherein the processing module is further configured to:

Determining a faulty node in the content distribution path; and,

And deleting the faulty node from the content distribution path, and using a downstream node of the faulty node as a downstream node of the upstream node of the faulty node.
The apparatus according to claim 22, wherein the processing module further comprises: a fourth determining unit, configured to determine, when the response message for the content distribution message sent by the first distribution target node is not received The first distribution target node is a faulty node in the content distribution path, wherein the first distribution target node is a distribution target node of any one of the content distribution paths;

a fifth determining unit, configured to determine a faulty node in the content distribution path according to the state traffic reported by the at least one distribution target node and the distribution log; and

a sixth determining unit, configured to: when receiving the distribution failure message sent by the second distribution target node, determine that the downstream node of the second distribution target node is a fault node in the content distribution path, where the The second distribution target node is any one of the distribution target nodes except the last distribution destination node in the content distribution path.
The device according to claim 22, wherein the processing module is further configured to:

Acquiring a recovery node in the content distribution path, the recovery node being a node deleted from the content distribution path due to a failure;

Adding the recovery node to the content distribution path, so that the recovery node acquires distribution content from an upstream node of the recovery node, wherein the recovery node is from an upstream node of the recovery node by an active request mode Get the distribution.
A computing node, comprising:

a receiving module, configured to receive a content distribution message, where the content distribution message includes a content distribution path and an identifier of the content to be distributed, wherein the content distribution path is configured by the distribution controller according to a physical networking topology of the cloud computing system, and the cloud computing a state traffic of each node in the system and a distribution target node list, where the content distribution path is a unidirectional chain distribution path formed by the first node and the at least one distribution target node, where the first node is in the system a complete node of the content to be distributed;

a processing module, configured to acquire, according to the identifier of the content to be distributed, metadata of the content to be distributed from a first node, where the metadata of the content to be distributed includes a digital signature, a block index, and a content of the content to be distributed Block hash value;

The processing module is further configured to perform content distribution processing according to the content distribution path and the metadata of the content to be distributed.
The computing node according to claim 25, wherein the content distribution message further includes a specified distribution rate of the distribution target node;

The processing module is specifically configured to perform content distribution processing according to the specified distribution rate according to the content distribution path and the metadata of the content to be distributed.
The computing node of claim 26, further comprising:

a sending module, configured to report a ready state to the distribution controller;

The receiving module is further configured to receive a content receiving and forwarding indication sent by the distribution controller.
The computing node of claim 25, wherein the processing module comprises:

a receiving unit, configured to receive a block sent by an upstream node;

a checking unit, configured to perform hash check on the block according to the block index and the block hash value in the metadata of the content to be distributed, and obtain a verification result;

And a sending unit, configured to save the block when the verification result is a verification success, and send the block to a downstream node in the content distribution path according to the specified distribution rate.
The computing node according to claim 28, wherein the sending unit is specifically configured to:

If the transmission fails, resend; and,

If the number of retransmission failures reaches a preset number of times, a distribution failure message is sent to the distribution controller to cause the distribution controller to re-establish the content distribution path according to the state of the downstream node.
A computing node according to claim 28 or 29, characterized in that

The processing module is further configured to: when receiving all the blocks corresponding to the content to be distributed, perform digital signature check on the to-be-distributed content according to the digital signature, and report the result of the digital signature check to the distribution controller .
A content distribution system, comprising the content distribution device of any of claims 16-24 and the computing node of any of claims 25-30.